Rural inventory

1. Spatial classification of the whole Territory of Serbia into urban and rural areas based on Corine Land Cover data.
2. Spatial interpolation of the MDTY values (for each vehicle category) from VCDs at the mid-points of the roads where VCD not exist. This will be done for each category of road, separately, by using the knn method (k-nearest neighbors). In doing so, the distances to the whole road section where AVC exist will be taken into account. Model will be trained based through the cross-validation procedure, by using the RMSE (Root Mean Squared Error) as a measure of performace. In this way, the vehicle activity for each road will be estimated.
3. Based on these estimates, total rural inventory for each vehicle category will be spatially dissagregated into the 5x5km cells propotionally, by taking the estimated MDTY and the lengths of corresponding road sections that lie within the cell into account.

Urban inventory

1. First, total inventory pollution will be spatially dissagregated based on the number of the vehicles registered in each municipality.
2. Next, estimated urban emissions in each municipality will be further spatialized into 5x5km cells based on the the lenght of the roads that lie in each cell of urban areas.

Highway inventory

• Total inventory for highway transport will be spatially dissagregated based on the MDTY values and the lenght of the corresponding highway section that lie in each cell.

Total inventory

• Total air pollution from the road transport in each cell will be estimated as the sum of rural, urban and highway inventory for each cell proportionally.